Конференция Ruggedize: технологический прорыв в роботизированном сельском хозяйстве

A new technical conference for AI and ag tech entrepreneurs, as well as investors, launched last week as 300 gathered for Ruggedize, a first-of-its-kind meeting, at Reservoir Farms, an ag tech incubator housed in a barn in a Salinas farm field. The conference setting, in a tent with a dirt floor, was a contrast of high tech and low tech, launching what Reservoir Farms CEO and event organizer Danny Bernstein called “the Golden Age of robotics in agriculture.” Influential tech companies including Google, Nvidia, Amazon, Yamaha and Robot Operating Systems showed off free or open-source resources for developers creating new solutions. They were joined by ag tech giant John Deere and successful ag tech innovators from Carbon Robotics, Bonsai Robotics, Verdant Robotics and others who showcased AI breakthroughs and business models. At the conference, John Deere announced it has committed $10 million over three years to support the Reservoir Farms incubator, which houses a dozen entrepreneurs in Salinas and more in satellite incubators in Santa Rosa and Arizona. John Deere allows Reservoir Farms' developers free use of its tractors, but does not receive any equity in return. A separate investment entity, Reservoir VC, aims for $50 million from investors to fund “rugged AI startups that solve one high-value task in ag and then compound to adjacent industries,” Bernstein said. Bernstein said the pool of tech resources is enabling innovation at the grassroots level. “Shared tooling, affordable edge compute, simulation, and open-source foundations now mean a two-person team can build a real company in weeks instead of years,” he said. “And that changes who gets to build,” he added. “When the foundational work is shared, the barrier to entry collapses, and the people solving the hardest problems in the field can finally build on each other's work instead of alone. That is how a golden age starts.” Democratizing the Tools Paige Bailey, of Google’s DeepMind, demoed Gemma 4’s capabilities that enable local compute, eliminating the need for a Wi-Fi connection for ag tech. “Pretty wild, eh? I think that this is something that I'm most proud of when it comes to truly democratizing AI,” said Bailey, a Texan-born engineer from a farming family outside of Waco. “Not everybody can afford a big cloud bill. Not everybody has access to Wi-Fi ... you can bring these models to the people to make them as powerful as you can. …The world gets to be a much more interesting place since all of this is open source ... or open models in terms of the Gemma models that you can download.” In her talk, “Building Rugged Autonomous Systems with Open Source,” Robot Operating System expert Kat Scott focused on building blocks anyone can use to build smart AI applications for unpredictable farming environments. ROS is used in many robotics classes. Ag Tech Expands Multi-Crop, Multi-Purpose Equipment While first-generation ag robots were typically focused on a particular crop or task, there has been a shift towards multi-tasking, multi-crop machines, experts said. After acquiring Blue River Technology in 2017 for $305 million, John Deere launched the AI-enabled sprayer’s core capabilities in its See & Spray product in 2021. By 2025, that product was used on more than 5 million acres of corn, soybean and cotton. The weed-sensing, precision ag technology “demonstrated an average herbicide savings of 59% on corn, soybean and cotton fields across the U.S,” said Sean Sundberg, John Deere’s business integration manager. That translates into an economic savings of $15.70/acre, according to an Iowa State study the company cites. See & Spray will expand into barley, canola, edible beans, peanuts, milo, sugar beets and wheat in the 2027 model year. A New Frontier: Precision Fertilizer Application At 86, Jorge Heraud, founder and former CEO of Blue River Technology, is on to his next startup, TerraBlaster, which recently raised $12.5 million for real-time NPK soil mapping. Its 0.15-acre resolution — roughly 16 times finer than a conventional 2.5-acre grid sample — increases efficiency. "We are quickly moving to a future where real-time measurements enable precise soil fertilization," Heraud said. USDA Engagement The USDA is also arcing toward the future of AI-enabled ag robotics, creating a nationwide imaging program, the Digital Ag Systems Hub (DASH). Steven Mirsky, director of digital agriculture at USDA's Agricultural Research Service, said, “This is an institutionalized effort across USDA to help curate ag image repositories, create this open access ag image repository, and also help with the development of end-to-end computer vision solution.” A closing panel on the second day of the conference provided a reality check on the ease and speed of progress. “Being able to deploy ruggedized robots into agricultural settings here in this part of the world, or anywhere else where we grow food, is really, really challenging,” said Ben Palone, senior director of automation and commercialization for Western Growers. But, he said, seeing the wide range of off-the-shelf developer resources was encouraging. “Founders can really start doing that development of technology, not worrying about having to build everything from the ground up,” he added. Bernstein echoed some of the same sentiments in his welcoming remarks. “Anything is possible right now. You really feel this idea that agriculture has a lot of limitations. Rugged environments have a lot of limitations, but when we think about who's collected here, we really begin to think that there's a lot more that's possible,” Bernstein said. Western Growers signed a three-year, $1.5 million partnership with Reservoir Farms in May. Reservoir Farms incubator company Sami Robotics is developing AI-enabled harvesters for broccoli, and Romaine and iceberg lettuce. The Ruggedize conference will be held again Aug. 25-26, 2027. The 2026 conference proceedings are all available online on Reservoir’s YouTube channel.